Choice of OptK

[Thomas Jones]

Additionally, I have a query regarding optK. I read in a previous post that you would recommend keeping optK the same when running CODEX2 on a single batch. I am doing this but have a less clear choice for an optK for all chromosomes. For example:

chr1: 

chr2:

chr22:

In this case would you recommend sticking with a single choice for optK that's a best fit for all, or should each chromosome be run using the best optK for the given chromosome? 

Thanks again for your help, 

Tom Jones 

The Francis Crick Institute Limited is a registered charity in England and Wales no. 1140062 and a company registered in England and Wales no. 06885462, with its registered office at 1 Midland Road London NW1 1AT

[Jiang, Yuchao]

HI Tom,

I am afraid I cannot give you a direct answer for this. It’s similar to the choice of number of clusters and is, in some sense, more like an art than science. If you do have some negative/positive controls, I think you can resort to them and additionally adopt extensive visualizations.

Note that if you have negative control samples, the choice of K is less influential because you only use the control samples to estimate the exon/target-specific factors.

Yuchao

On Aug 20, 2020, at 10:16 AM, Thomas Jones <thomas...@crick.ac.uk> wrote:

Additionally, I have a query regarding optK. I read in a previous post that you would recommend keeping optK the same when running CODEX2 on a single batch. I am doing this but have a less clear choice for an optK for all chromosomes. For example:

chr1: 

<image.png>

chr2:
<image.png>

chr22:

<image.png>

In this case would you recommend sticking with a single choice for optK that's a best fit for all, or should each chromosome be run using the best optK for the given chromosome? 

Thanks again for your help, 

Tom Jones 

The Francis Crick Institute Limited is a registered charity in England and Wales no. 1140062 and a company registered in England and Wales no. 06885462, with its registered office at 1 Midland Road London NW1 1AT

-- 
You received this message because you are subscribed to the Google Groups "CODEX/CODEX2 for CNV detection" group.
To unsubscribe from this group and stop receiving emails from it, send an email to codex2+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/codex2/VI1PR0502MB40960F84C84B1D3AA660251BC95A0%40VI1PR0502MB4096.eurprd05.prod.outlook.com.